RETRACTED ARTICLE: A novel panoptic segmentation model for lung tumor prediction using deep learning approaches

Clinicians can use X-ray image segmentation to help them analyze lung status and spot lung illnesses. In X-ray imaging systems, it could be required to develop rapid and novel segmentation methods without using cloud computing. When used on cutting-edge gadgets like X-ray imaging equipment for the lungs, this work investigates portable deep learning techniques for lung image segmentation. This research proposes a revolutionary scalable attention mechanism \(({\text{s-AM}})\) $(\text{s-AM})$ technique. On a collection of 360-degree X-ray images of the lungs, the suggested approach and many other lightweight deep learning methods were trained. Several accuracy criteria were used to analyze and compare these quick techniques. The trained model for the suggested scalable attention mechanism \({\text{s-AM}}\) $\text{s-AM}$ technique and it outperformed other methods in terms of IoU is 0.90, Dice is 0.92, HD is 4.9, VOE is 0.19 and RVD is 0.25. This work demonstrates that the lung X-ray image segmentation approach that was suggested only needs a modest amount of memory storage and yet performs comparably. The approach may be used on edge devices, and it might help physicians by streamlining their everyday tasks and enhancing the accuracy of their analysis.